To improve the potential of dose escalation we shall develop methods to deliver dose more precisely, with emphasis on reducing and correcting for organ motion, a principal factor that limits the accurate treatment of the lung and prostate. Since motion in the two sites are caused by different mechanisms, two strategies are proposed. In the lung, the aim is to immobilize the target with deep inspiration breath-hold (DIBH) an dverify its location at each treatment. Thus the high dose volume of normal lung may be minimized due to the reduced apertures an ddecreased lung density. Initially, the effect of respiration on tumor motion will be measured in a cohort of patients with DIBH spiral CT scans, and patient eligibility criteria will be established. Then, we shall develop and implement DIBH treatments with spirometric monitoring of inspiration level. The target positioning will be verified by comparing the location of chest wall and diaphragm on electronic portal images and digitally reconstructed radiographs. The potential benefits will be assessed by comparing treatment plans for DIBH with currently practiced free-breathing treatments. In this pelvis, variability in bladder and rectal filling causesinterfractinal motion. We propose a two-pronged strategy: the patients are planned using the average target position obtained from two or more CT scans, and the aperture margins designed from a generic data set to account for the residual fluctuations of the organs about their mean positions. The generic organ motion data will be measured in a cohort of patients using CT scans acquied at 4 points in time during the treatment course. We shall develop computer tools for optimizing treatment apertures using the generic motion and setup error data, and plan evaluation tools for predicting the variability of dose-volume histograms due to organ motion and setup errors. In addition, we will study CT guided therapy, with a CT scanner located in the treatment room. Patients are scanned prior to treatment, and the target position corrected (if necessary) for the treatment. This task will involve identifying those patients wo have highly mobile prostate by using CT scans acquired prior to or during the early phase of therapy. The initial trial will be on 50 patients for whom the setup is adjusted to register the center-of-mass of the prostate to the isocenter for each treatment fraction. In a later phase, a change in the field shap will also be studied. Finally, the value of such procedures will be assessed with dosimetric studies.